ainoya.dev

The capture screen in Ayumi, my voice-journaling app, sends a recording to a transcription model and shows the text back.

I have recently been building two apps for myself: Ayumi, a journaling app, and Kitae, a workout tracker.

import PRReviewSimulator from '../../..

The other day I watched an LLM agent build a working API server.

I built the documentation site for Pokke using Starlight, which is available at docs.pokke.femto-cloud.com.

I recently added a very simple, yet effective feature to pokke, a personal bookmarking service I am developing.

I recently created a simple digital toy called Emoji Shower.

I’ve been trying out the voice input feature in Antigravity (specifically testing the UI in version 1.15.8), and it sparked a realization about how we interact with voice UIs.

In a previous post, I introduced pokke, a minimalist bookmarking service I built for my own use.

Cursor can report Invalid OAuth error response when a custom Model Context Protocol (MCP) server returns a plain-text 404 from its OAuth metadata endpoint.

Like many developers, I relied on Pocket for years to save articles for later reading.

Introduction: The "Millionaire" Philosophy In the Japanese software engineering community, there is a concept known as "Millionaire Programming" (富豪的プログラミング).

Jules and Antigravity overlap as coding agents, but I do not use them for the same job.

A reverse centaur is not a person strengthened by a machine.

Around November 2025, a specific dance challenge began to go viral in Japan.

Browser MediaRecorder output can produce WebM files whose duration is missing or reported as Infinity.

Recently, the Model Context Protocol (MCP) has been gaining significant attention in the context of AI agent utilization.

Astro Cactus ships with dynamic Open Graph (OGP) images: for each post, the theme renders a title card using Satori (to SVG) and Resvg (to PNG).

I recently found myself architecting a seamless experience for an AI chat application.

The era of AI-powered software development is firmly upon us.

From Marble to Model: The AI Sculpting Process Creating something with generative AI often feels a lot like carving a Buddha statue out of a block of wood.

This YouTube video is licensed under CC BY.

When using generative AI to process large contexts, such as PDFs, the output may sometimes be cut off.

Recently, I've been experimenting with using Large Language Models (LLMs) for personal knowledge management.

LocalStack normally gives each developer a separate local KMS state.

The Rise of Generative AI and the Information Explosion The advent of generative AI has ushered in an era of unprecedented information output.

Developing a Web UI for Controlling LLM JSON Output Introduction Controlling LLM (Large Language Model) output to get the desired results often requires meticulous specification of output formats and content.

It's become standard practice to automate and codify product operations.

I've been using tags to organize and connect my notes locally.

Rendering Markdown Images in LLM Apps?

import Twitter from '../../..

I previously used DOM Distiller for content extraction in my project cloudflare-dom-distiller.

Introduction For my personal and professional needs, I rely on two powerful tools: Pocket for collecting and storing articles, and Obsidian for organizing and synthesizing my thoughts.

Recently, I encountered a challenge while working with machine learning models on Amazon EC2.

Introduction Maintaining and reinforcing a company's culture and values among employees can be a challenging task.

Securely managing sensitive information such as API keys in web services is crucial.

In a previous exploration, I utilized the openapi-generator tool to create a Java client for the OpenAI API.

Introduction A startup that was developing a personal asset management app recently ceased operations.

In the unsupervised learning approach of SimCSE, two identical sentences are vectorized using the Transformer encoder to train a sentence embedding model.

In pursuit of an efficient development workflow for my personal blog, ainoya.dev, I experimented with leveraging the capabilities of VSCode's DevContainer feature.

Introduction In this post, I'll explain a common pattern in chatbot conversations involving function calling within the backend during interactions with an OpenAI Assistant API.

When managing tickets in JIRA and managing code in GitHub, the development workflow often involves the following steps: Create a Git branch named after the JIRA ticket number.

Astro, a powerful framework for building modern websites, offers efficient static build capabilities.

Introduction In this article, I'll introduce an example of implementing secure service-to-service communication using Envoy.

Early in the previous year, following the launch of OpenAI's GPT-3.5, I initiated a project to improve existing tools' user experience using the OpenAI API.

Last year in November, I took on the task of developing a Java API client for OpenAI's ChatGPT Assistant API, as there was a notable absence of such a tool for Java developers.

In the world of data analytics, Looker has become an indispensable tool for many teams.

Introduction In the dynamic world of software engineering, the integration of efficient tools can significantly streamline the workflow.

Enhancing Home Security with a Custom-Built Motion Detection Alert System using SwitchBot and Cloudflare Workers As smart home technologies evolve, the need for more sophisticated monitoring systems becomes apparent.

I wanted to share a small update about some changes I've made to my website recently.